The dic(9;20)(p13.2;q11.2) is reported to be present in ∼2% of childhood B-cell precursor acute lymphoblastic leukemia (BCP ALL). However, it easily escapes detection by G-banding analysis and its true prevalence is hence unknown. We performed interphase fluorescence in situ hybridization analyses-in a three-step manner-using probes for: (i) CDKN2A at 9p21, (ii) 20p and 20q subtelomeres and (iii) cen9 and cen20. Out of 1033 BCP ALLs diagnosed from 2001 to 2006, 533 were analyzed; 16% (84/533) displayed 9p21 deletions, of which 30% (25/84) had dic(9;20). Thus, dic(9;20)-positivity was found in 4.7% (25/533), making it the third most common genetic subgroup after high hyperdiploidy and t(12;21)(p13;q22). The dic(9;20) was associated with a female predominance and an age peak at 3 years; 18/25 (72%) were allocated to non-standard risk treatment at diagnosis. Including cases detected by G-banding alone, 29 dic(9;20)-positive cases were treated according to the NOPHO ALL 2000 protocol. Relapses occurred in 24% (7/29) resulting in a 5-year event-free survival of 0.69, which was significantly worse than for t(12;21) (0.87; P=0.002) and high hyperdiploidy (0.82; P=0.04). We conclude that dic(9;20) is twice as common as previously surmised, with many cases going undetected by G-banding analysis, and that dic(9;20) should be considered a non-standard risk abnormality.